TIDo: Source-free Task Incremental Learning in Non-stationary Environments

TL;DR

This paper introduces TIDo, a source-free, one-shot task incremental learning method capable of adapting to non-stationary environments by minimizing adversarial discrepancy and using Gaussian prototypes, without storing past data.

Contribution

The proposed approach enables incremental learning of source and target tasks in non-stationary environments without storing past data, using adversarial discrepancy minimization and Gaussian prototypes.

Findings

Achieved improved performance over state-of-the-art methods.

Effectively handles non-stationary source and target tasks.

Reduces catastrophic forgetting with distillation loss.

Abstract

This work presents an incremental learning approach for autonomous agents to learn new tasks in a non-stationary environment. Updating a DNN model-based agent to learn new target tasks requires us to store past training data and needs a large labeled target task dataset. Few-shot task incremental learning methods overcome the limitation of labeled target datasets by adapting trained models to learn private target classes using a few labeled representatives and a large unlabeled target dataset. However, the methods assume that the source and target tasks are stationary. We propose a one-shot task incremental learning approach that can adapt to non-stationary source and target tasks. Our approach minimizes adversarial discrepancy between the model's feature space and incoming incremental data to learn an updated hypothesis. We also use distillation loss to reduce catastrophic forgetting of previously learned tasks. Finally, we use Gaussian prototypes to generate exemplar instances eliminating the need to store past training data. Unlike current work in task incremental learning, our model can learn both source and target task updates incrementally. We evaluate our method on various problem settings for incremental object detection and disease prediction model update. We evaluate our approach by measuring the performance of shared class and target private class prediction. Our results show that our approach achieved improved performance compared to existing state-of-the-art task incremental learning methods.